2D Drawing Visualization Framework for Applying Projection-Based Augmented Reality in a Panelized Construction Manufacturing Facility: Proof of Concept

Product quality is recognized as a major benefit in industrialized construction because of the utilization of machines in manufacturing facilities. However, manual work continues to exist in complex assembly projects, which can potentially cause quality issues. To address such quality issues in industrialized manufacturing processes, projection-based augmented reality (AR) techniques have been applied in other industries (e.g., manufacturing), but on a smaller scale and at a short distance. To apply projection-based AR in the construction manufacturing facility, this paper proposes a framework that enables the user to perform a vision-based projection alignment using a projector and camera. The designated projection area determined by the user is marked out, the coordinates of which are then computed through segmentation and object-detection algorithms. After the acquisition of the markers' coordinates is complete, a two-dimensional (2D) image is overlaid on the surface of the designated area using a projector by computing transformation matrix for the projection. To evaluate the potential performance in a field setting, the offset distances between the four corners of the projection boundary and the center of the markers are measured in various environments, such as at distances ranging from 5 to 8 m, and with different illumination conditions (i.e., low and high brightness) as well as in a manufacturing shop. The results indicate that average offset distances in all the experimental conditions are shorter than the factory tolerance level of 6.35 mm. Also, the statistical analysis reveals that both the distance and the illumination are significant factors affecting the projection alignment performance. This study provides a novel approach for visualizing vital information within a user's field of view during the manufacturing processes at shops and offers considerations for implementing projection-based AR in practice.